Transcript Second

Intermediate Goods, Weak
Links, and
Superstars: A Theory of
Economic Development
Charles I. Jones
Department of Economics, U.C.
Berkeley and NBER
February 13, 2008– Version 2.0
Charles I. Jones

Department of
Economics, U.C.
Berkeley
Currently visiting the
Stanford GSB
 Email:
 [email protected].
edu
 http://www.econ.berke
ley.edu/˜chad
The Main Contents of This Paper
 This
paper try to explain the enormous income
difference between the richest countries and
the poorest countries, and proposes that
linkages, complementarity, and superstar
effects are at the heart of the explanation.
 First, linkages between firms through
intermediate goods deliver a multiplier similar to
the one associated with capital accumulation in
a neoclassical growth model.
The Main Contents of This Paper
 Second,
just as a chain is only as strong as its
weakest link, problems at any point in a
production chain can reduce output
substantially if inputs enter production in a
complementary fashion.
 Finally, the high elasticity of substitution
associated with final consumption delivers a
superstar effect: GDP depends
disproportionately on the highest levels of
productivity in the economy.
The Main Contents of This Paper

This paper builds a model with links across sectors,
complementary inputs, and highly substitutable
consumption, and shows that it can easily generate 50
fold aggregate income differences with plausible
differences in distortions and productivity across
countries.
 The political economy approach explains why
resources are misallocated (see Acemoglu and
Johnson (2005) and Acemoglu and Robinson (2005));
the approach here explains why misallocations lead to
large income differences. Clearly, both steps are
needed to understand development.
The Framework







Part 2 discusses briefly the key mechanisms at work
in this paper: complementarity, substitution, and
linkages through intermediate goods
Part 3 applies the basic discussion in part 2 to
construct a theory of economic development.
Part 4 analysis the weak link effect.
Part 5 analysis the superstar effect with plausible
differences in distortions.
Part 6 attempts to quantify the mechanisms at work in
our theory.
Part 7 make some expansion of this model.
Part 8 is the conclusion.
2. LINKAGES, COMPLEMENTARITY, AND
SUPERSTARS

Linkages between sectors through intermediate goods
deliver a multiplier very much like the multiplier
associated with capital in the neoclassical growth
model. More capital leads to more output, which in
turn leads to more capital. This virtuous circle shows
up mathematically as a geometric series which sums
to a multiplier of 1/(1− α), if α is capital’s share of
overall revenue.
 Intermediate goods generate this same kind of
multiplier. In the model below, this multiplier depends
on 1/(1− σ) ,where σ is the share of intermediate
inputs in gross output.
 2.1
Linkages through Intermediate Goods.
 In the model, the overall multiplier (for example,
on productivity) is the product of the
intermediate goods and capital multipliers:
1/(1− σ) × 1/(1− α), = 2 × 3/2 = 3.
 Combining a neoclassical story of capital
accumulation with a standard treatment of
intermediate goods therefore delivers a very
powerful engine for explaining income
differences across countries. Related insights
pervade the older development literature but
have not had a large influence on modern
growth theory.
 2.2
The Role of Complementarity.
 Because of complementarity, the development
problem may be hard to solve. In any
production process, there are ten things that
can go wrong that will sharply reduce the value
of production. In rich countries, there are
enough substitution possibilities that these
things do not often go wrong. In poor countries,
on the other hand, any one of several problems
can doom a project.
 Linkages through intermediate goods provide a
large multiplier, while complementarity means
that there is typically not a single magic bullet
that can exploit this multiplier.
2.3 An Example of
Complementarity:
Socks!
2.4. Modeling Complementarity and Substitution
A natural way to incorporate varying degrees of
complementarity and substitution is with a constant
elasticity of substitution (CES) function. For example,
suppose:
Let z i denote a firm’s purchases of the i-th input, and
assume a continuum of intermediate inputs are
necessary for production.
The elasticity of substitution among these activities is:
1/(1   )
but this (or its inverse) could easily be
called an elasticity of complementarity instead.
For intermediate inputs, it is plausible to assume:
  0 , So the elasticity of substitution is less than
one:
1
 1,
1 
for   0
The limiting case where:    . This means a
perfect complementarity, is also means that inputs
cannot be substituted. In this case the CES function
converges to the minimum function, so output is equal to
the smallest of the inputs z
What would the CES function converges to?
when:
  1&1;   0;   ;   
?
Proof : A simpler case
Suppose that

1 1
 1/
2 2
Y  ( z   z )
then l i mY  1 z1   2 z2 ;
 1
1
l i mY  1
( 即调和平均)

1
 1
1 z1   2 z2
1

 
then l i mY  l i m( z   z )  l i mexp  ln(1 z1   2 z2 ) 
 0
 0
 0



1 1
 1/
2 2
1
2
1  2 1  2
1
2
 z ln z1   z ln z2 
 l i mexp 
z


 0
1 z1   z



1 1

2 2

2 2
z
then see t he case   - 
if x1  x2
l i mY  l i m(1 z1   2 z2 )1/ < l i m(1 z1   2 z1 )1/
 - 
 - 
 - 
 l i m(1   2 )1/ z1  z1
 - 
l i mY  l i m(1 z1   2 z2 )1/ > l i m(1 z1 )1/  l i m(1 )1/ z1  z1
 - 
 - 
 - 
 - 
vice ver sa, so we get l i mY  min  zi  ( 夹逼准则)
 - 
Goi ng i n t he ot her di r ect i on, i f   
We have l i mY  max  zi 
 +
3. SETTING UP THE MODEL
A continuum of goods indexed on the unit interval
by i are produced in this economy using a relatively
standard CobbDouglas
production function:
The GDP in this economy is:
These final consumption goods aggregate up with an
elasticity of substitution greater than one, because:
1
 1,
1
for 0    1
intermediate inputs combine with an elasticity of
substitution less than one. This is the key place where
“weak links” enter the model:
1
the elasticity of substitution:
 1,
1 
for   0
This aggregate intermediate good is what gets used
by the various sectors of the economy. To keep the
model simple and tractable, we assume that the
same combination of intermediate goods is used to
produce each variety
Attention: 要素如果被用来做中间产品进行生产,则它们
是互补的,不可替代;如果是用来做为最终消费品,则具有
很高的替代性.
The constraints of the model:
Finally, preferences are standard:
with u( C) >0 and u( C) <0
The share of produced goods in the production of
good i is therefore:  (1   )  

The parameter
measures the importance of
linkages in our economy. if   0 the productivity
of physical and human capital in each variety
depends only on Ai and is independent of the rest of
the economy. If σ>0, low productivity in one sector
feeds back into the others.
4. A SYMMETRIC ALLOCATION OF RESOURCES
PROPOSITION 1 Proof.
step1: According t o def i ni t on 1:
Yi = Ai ( K i H i1 )1 X i  Ai ( K  H 1 )1 X   Ai m
wher e m = (K  H 1- )1- X 
step2: Using (5) X =


1
0

1
0
zi di
( zAi m)  di


1/ 
1/ 



1
0

1
0
( zYi )  di
Ai di


1/ 
1/ 
zm
 S  zm  S  z(K  H 1- )1- X 
 X  S  z(K  H 1- )1- X   X   ( S  z )
 m  (K  H 1- )1- X   ( S  z )

1
K  H 1-

1
(K  H 1- )
step3: Using (4) Y =
1/ 



   (Yi  zYi ) di 
 0

1


1
0
Ai di

1/ 

1
0
ci di

1/ 

1/ 


   (Yi  zi ) di 
 0

1

1/ 


   ( Ai m  Ai mz ) di 
 0

1
(1  z )m  S (1  z )m  S (1  z )( S  z )
  ( z )( S 1 S  )
1
1
K  H 1-
where  ( z )  ((1  z )1 z  )
1
1

1
K  H 1-
1
when   ,  = ,  =1
2
1
lim S 
 - 
1/ 2
S     Ai di  min  Ai 
 0


1
Aggregate TFP then depends crucially on the smallest
level of TFP across the sectors of the economy—that is,
on the weakest link.
Second, Total factor productivity involves a multiplier,
the exponent:
1
1
 1
Finally, consider the role of:
1
1
 1
 ( z )  ((1  z ) z )
1
 ( z )
1
 
1  1
1




(1


)(1

z
)
z


(1

z
)
z
)

z
1 
 ( z )
let
0
z
 (1   )(1  z )  z    (1  z )1 z  1

z


 z 
1 1 z
this turns out to be the optimal amount of gross
output to spend on intermediate goods. Departures
from this optimal amount will reduce TFP.
5. A COMPETITIVE EQUILIBRIUMWITHWEDGES
Letting the final output good be the numeraire, these
problems are described below.
Household Problem:
Solving for the household problem:
Let Hˆ = u (C )  q (t )  rV  w H  T  C
t
t
t
t
t
t

Hˆ
 u (Ct )  q (t )  0  u (Ct )  q (t )
Ct
 q (t )  u (Ct )
Hˆ
 q (t ) rt   q (t )  q (t )  q (t )  (  rt ) q (t )
Vt
 (  rt )u (Ct )
u (Ct )
Hence, (  rt ) 
, if the utility function is
u (Ct )
constant-relative-risk-aversion, then we have:
C  r

0 r
C

Final Sector Problem:
Intermediate Sector Problem:
Variety i’s Problem:
5.3. Solving for the Competitive Equilibrium
Proof: see appendix
First, the multiplier
1
associated with intermediate
1
goods appears in exactly the same way as in the
symmetric allocation, and for the same reason.
Second, the tax wedges affect output through TFP.
Therefore, this proposition illustrates a very important
result found elsewhere in the macro literature: the
misallocation of resources at the microeconomic level
often shows up as a reduction in TFP at the
macroeconomic level.
Finally,
suppose   , since

1 
 1





1

lim Q1 Q  Q1  1
 - 
1


d
i

0 A (1   )


i
i


the second term is the harmonic mean of the
tax adjusted productivity
Disasterously low productivity in a single variety is fatal
in the symmetric allocation, but not in the equilibrium
allocation. Why not?
The reason is that the equilibrium allocation is able to
strengthen weak links by allocating more resources to
activities with low productivity.
Further more:
suppose   ,  =1,  =1/2


since
 1;
 
1 
1-
lim Q1 Q  lim Q1 Q
 - 
 1
 - 


1-
1/ 2


1 

1
2 


ma
x
A



i


1
 1

di



 0 Ai (1   i )

only the good with the highest productivity
will be consumed: the arithmetic mean gets
replaced by the max  Ai  -----a superstar effect.
5.4. The Steady State
Proof: step 1
 1/(1 )
Y =  ( )(Q1 Q
)
K  H 1
Y
Y
K

ke 
H
h L
h L
 1/(1 )
A= ( )(Q1 Q
)
let y e 
then y e = A( k e )
 (1- )
1 
Step 2: Using (A5) K =
Y
r+ 1   (1   )
 (1- )
1 
Let c =
r+ 1   (1   )
 k  cy  k e  cy e
According to step 1: y e = A( k e )
 k e  cA( k e )  k e = (cA)1/(1 )
 y e = A(k e )  A(cA) /(1 )  A1/(1 ) c /(1 )
 y  A1/(1 ) c /(1 ) h
  ( )(Q Q )
1
 1/(1 )

  1 ( )(Q Q )
1


1
1
1 1

1 /(1 )
  (1- )

1 
 r+ 1   (1   ) 


1/(1 )
  (1- ) 
 r+ 
h , (r= )
 /(1 )
h
5.5. Symmetric Tax Wedges
5.6. Random Tax Wedges
6. DEVELOPMENT ACCOUNTING
Attention:
The point of these numerical examples is
not to claim that they precisely
characterize the nature of income
differences between the richest and
poorest countries. Rather, they illustrate
that these mechanisms are able to
generate enormous differences using
relatively conservative choices of
parameter values.
7. THE CUMULATIVE EFFECT OF REFORMS
8. CONCLUSION
First, forces of complementarity and substitution lead
extreme sectors—both the superstars and weak links—
to play disproportionate roles. Income in the rich
countries may depend primarily on superstars, while
income in the poor countries may depend on the
weakest links.
Second, The presence of intermediate goods leads to a
multiplier that depends on the share of intermediate
goods in gross output.
Finally, An important channel for future research
concerns the role of intermediate goods.